My bench power supply..

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
After some experimenting and studying some PS schematics ,ive
decided to build my own, want something so reliable to test my
projects..

the following is my own design (thanks to many pals in here,
they really helped me alot!)..
its a 1.5A regulated providing +-1.25 upto 35V using the LM317,
LM337 reuglators (can be substituted with more current handlers
like the LM350, LM138..etc)

i need to hear all of ur comments, suggestions on the main schematic and how to improve the output or to provide good
protection...etc

waiting ur valuable goods!
thanks..
 

Attachments

  • ps.jpg
    ps.jpg
    49.2 KB · Views: 1,526
You won't be able to output very high currents at low voltages.
For instance, if you tried to draw 1.5A at 1.25V, the 317 would
have to dissipate around 50W. This is far above their rating. I
don't remember exactly what the power rating is, but it is at
most 20W, probably less. Besides, the 317 does not have very
good thermal conductivity from junction to case, so in practice
it is very difficult to get even near their rated maximum power
without gigantic heat sinks.
 
My Bench Supply - I will draw a schematic - some time

I have a bench supply I made it myself.
I took what I had at home, and got it together.
My own circuit design.
--------------------------------------
It uses a LM385 as reference (set at +5.00V).
A pot can adjust Ref output 0.0-5.0 V
This is buffered and mirrored to a negative 0.0-5.0 V reference.

It uses one OP741 and one LM324, quad OP.
One 7808 and and a dicrete OP to mirror +8.0 V to -8.0 V
It uses BD241 BD242 as output transistors,
controlled by 2 OP amps in LM324.
With some homemade heatsinks of 2 mm Aluminium plates.
--------------------------------------
I have +5.0 +8.0 0V -8.0 voltages
and a tracking dual adjustable +- 0.0-20.0 V outputs.
I have Corse/Fine pot for the adjustable +-0-20V,
and also a pot to trim the negative +-5% to match positive exactly.
Output current is like 2x 0.20-0.30A
I have used 15-0-15V 0.5A trafo.

I am proud to be able to put out voltages down to
plus/minus Zero 0.00V.
--------------------------------------
To sum:
+8.0 +5.0 0.0 -8.0 Volt
Adjustable dual Track: +- 0.0 to 20 Volt, 0.25A
--------------------------------------
I will draw a schematic - some time.
And put it into my FTP-Server, along with my collection


/halo - do it yourself - get it as you like to have it :)
 
if it's truly a bench supply

then you want to be able to get down to "0" volts which means you need to bias the adjust pins of the regulators -- you need to pull the LM317 1.25 volts below ground -- since these devices are so cheap (about $0.40 in the US) just use one with a minimal value Radjust to derive the bias.

Nat Semi has an apnote on this.
 
Sam9

A- If you find you come up short on current at voltages, you can use parrallel the regulators (2 or each). I tried this once when I needed more current that the LM317 could provide. Therewas probably sub-optimal about my approach, but it worked. They ran cooler, too.

B- On the other hand, if you want a firm upper limit of 1.5A you can put a series pass limiter after the the regulator(s). If you need a reference on that, G.R. Slone shows a basic one in his intoductory book on electronics.

If all you need to do is confirm amplifier function and set the initial bias, you don't need a "full power" PS. In fact, it may keep you from damaging the amp during trouble shooting (especially if current limited).
 
I'm building something very similar at the moment :)

You won't get 1.5A out at +-35V because the capacitors will discharge below the dropout voltage of the regulators. You should be able to get 1.5A at 26V out, judging by Duncan's PSU designer. (available at http://www.duncanamps.com/psud2/ )

Things to remember:
* regulators have a drop-out voltage, which is a miminum voltage difference between their Vin and Vout terminals for them to work. Typical values are about 3 volts, meaning that the output voltage needs to be about 3 volts less than the minimum voltage across the input caps.
* the lower the output voltage, the more power the regulator is dissipating at a given output current, as Christer pointed out
* the higher the output current, the lower the minimum voltage across the input caps.

Should be a nice power supply. Post pics when you've finished!
 
I wanted to make a split rail adjustable powersupply, except the only transformers I have are a 32volt 6amp one, and a 24volt 10 amp one.... and the first will exceed the max input of adjustable regulators..... I only just thought of using 1 transformer per rail, but always limiting the 32volt one to the same as the 24volt one.... But.... That would be one PRETTY big powersupply.... and if I wanted to get full power into say a +-28 volts, thats a PRETTY BIG drop for the regulator on the 32volt transformer, and thats ALOT of heatsinking needed.. :D lol
 
sam9 said:
A- If you find you come up short on current at voltages, you can use parrallel the regulators (2 or each). I tried this once when I needed more current that the LM317 could provide. Therewas probably sub-optimal about my approach, but it worked. They ran cooler, too.
(especially if current limited).

this is a point i really want to achieve (getting more current or even to make the regulators provides thier max limits cooler)
but itsnt that obvious how to parallel em! , can u explain more?




Sud said:
* the lower the output voltage, the more power the regulator is dissipating at a given output current, as Christer pointed out
* the higher the output current, the lower the minimum voltage across the input caps.[/B]


those are a must to know notes about the regulators of course
but is there anyway that we can make the regulators less to suffer in the power dissipation issue?? "sure by using agood cooling methode (heatsinks and fans which i already took into consid.)" but any other ways?
 
SkinnyBoy said:
I wanted to make a split rail adjustable powersupply, except the only transformers I have are a 32volt 6amp one, and a 24volt 10 amp one.... and the first will exceed the max input of adjustable regulators..... I only just thought of using 1 transformer per rail, but always limiting the 32volt one to the same as the 24volt one.... But.... That would be one PRETTY big powersupply.... and if I wanted to get full power into say a +-28 volts, thats a PRETTY BIG drop for the regulator on the 32volt transformer, and thats ALOT of heatsinking needed.. :D lol

Choke-input? ;)

Yeah, throw on some 2N3055s for some extra current, pull 5A through that motha.

- FWB the 24V, should get around 35VDC, derate to around 6-7A after rectification... FWB + choke input on the 32V tranny, current needn't be derated as much, but DC voltage is about .9 times the AC value, so they'll end up almost the same voltage and current.

Tim
 
How to wire regs parallel

When I did it it was simple: Take two LM317s, Connect both input leads and then both output leads to the same place as you would if it were only one reg. With reference to your schematic, each reg has its own R1. Also each has it's own R3 to ground. This is a little tricky in your case you need to find dual ganged pots but at least they don't need to be the expensive variety. (I was using a dual switch with specific resistor values.)

Although it may look tempting, I don't think you should use a common single pot - if there were to be any imbalance between regs, one of them might see a reverse voltage on the adjustment pin - probably a bad thing. (PS: watch the pinouts - in and out are reversed between the pos and neg versions. Getting one bassakwards can be exciting:redhot:
 
wiring regulators in parallel

I suggest you place a low resistance at the output of each regulator and connect the ends of the resistors together. A resistance of 0.1 Ohms would be adequate.

Also, upon second look, I suggest placing a 1uF to 10uF tantalum capacitor in parallel with R3 and R4 - it will help improve line requlation.
 
Re: wiring regulators in parallel

Sud said:
I suggest you place a low resistance at the output of each regulator and connect the ends of the resistors together. A resistance of 0.1 Ohms would be adequate.

Also, upon second look, I suggest placing a 1uF to 10uF tantalum capacitor in parallel with R3 and R4 - it will help improve line requlation.

Hi,

With respect, the usual reason for bypassing the resistors from the adjust pins to ground (R3 & R4) with a suitable capacitor, is to increase *ripple rejection*, and I have not noticed any improvement as far as the suggested *line regulation* goes in cases such as this. :)

Regards,
 
Higher voltage regulators

the TL78C from Texas Instruments is a little more expensive than the LM317, but will stand off 125 volts -- you will have to outboard a PASS transistor as the '783 will only take 700ma.

The easiest way to regulate a higher voltage is to "roll your own". Here's a high voltage <b>negative</b> regulator you can control with a potentiometer and an LM317 used as an adjustable reference. The same design can be adapted to use of a PNP transistor for a positive regulator. It's borrowed, loosely, from Horowitz and Hill and doesn't show some of the compensation, etc. which should be put in. The voltage sense pot actually characterizes the gain of the circuit. Disadvantage of an LM317 vs a DAC is that you can't get down to "zero" volts unless you bias the adjust pin below the reference voltage -- but it's simple and cheap:
 

Attachments

  • easy_neg_reg.jpg
    easy_neg_reg.jpg
    10.2 KB · Views: 560
Re: wiring regulators in parallel

sam9 said:
When I did it it was simple: Take two LM317s, Connect both input leads and then both output leads to the same place as you would if it were only one reg. With reference to your schematic, each reg has its own R1. Also each has it's own R3 to ground. This is a little tricky in your case you need to find dual ganged pots but at least they don't need to be the expensive variety. (I was using a dual switch with specific resistor values.)

Although it may look tempting, I don't think you should use a common single pot - if there were to be any imbalance between regs, one of them might see a reverse voltage on the adjustment pin - probably a bad thing. (PS: watch the pinouts - in and out are reversed between the pos and neg versions. Getting one bassakwards can be exciting:redhot:

u mean that every regulator (total of 4) should has its own
pot. ,so how can i adjust the pots. to provide a spec. volt. per rail
cant i use only one so it gonna be more simple and logic??



Sud said:
I suggest you place a low resistance at the output of each regulator and connect the ends of the resistors together. A resistance of 0.1 Ohms would be adequate.

Also, upon second look, I suggest placing a 1uF to 10uF tantalum capacitor in parallel with R3 and R4 - it will help improve line requlation.

ive seen a schematic once that places a 33ohm thermal resistor
before the regulator, dont know what its used for , but here u
said to place a v.small one (0.1ohm) at the output...!
can u explain why??


ps. ill get all of ur valuable comments & corrections and then
redraw the schematic again and re-post it of course :)
 
the resistor before the regulator

is current sense for the regulator and pass transistor combo -- as the current drawn increases (and the voltage drop across the transistor increases) it reduces the amount of work which the regulator has to perform and shifts this function to the transistor.

if you tie two LM317's together as described you can get some nice oscillations -- in fact, getting an LM317 to oscillate isn't difficult at all. At any rate, it's using a more expnesive part to do a less expensive job.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.